Model: SoLU Model: 8 Layers, 4096 Neurons per Layer

Dataset: 80% C4 (Web Text) and 20% Python Code

Neuron 3568 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 1.0295. Min Range: -1.0295

Max Act: 1.0295. Min Act: -0.0001

Data Index: 1548900 (Python Code)

Max Activating Token Index: 916

Text #1

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.8686. Min Act: -0.0001

Data Index: 521394 (C4 (Web Text))

Max Activating Token Index: 187

Text #2

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.8699. Min Act: -0.0001

Data Index: 824468 (C4 (Web Text))

Max Activating Token Index: 194

Text #3

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.8906. Min Act: -0.0001

Data Index: 1651603 (Python Code)

Max Activating Token Index: 632

Text #4

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.8190. Min Act: -0.0001

Data Index: 631808 (C4 (Web Text))

Max Activating Token Index: 87

Text #5

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.8388. Min Act: -0.0001

Data Index: 399709 (C4 (Web Text))

Max Activating Token Index: 59

Text #6

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.8324. Min Act: -0.0001

Data Index: 1539310 (Python Code)

Max Activating Token Index: 145

Text #7

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.8246. Min Act: -0.0001

Data Index: 1400597 (Python Code)

Max Activating Token Index: 486

Text #8

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.7943. Min Act: -0.0001

Data Index: 882740 (C4 (Web Text))

Max Activating Token Index: 83

Text #9

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.8421. Min Act: -0.0001

Data Index: 1332450 (C4 (Web Text))

Max Activating Token Index: 238

Text #10

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.7889. Min Act: -0.0001

Data Index: 272381 (C4 (Web Text))

Max Activating Token Index: 523

Text #11

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.8012. Min Act: -0.0001

Data Index: 1558750 (Python Code)

Max Activating Token Index: 718

Text #12

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.7743. Min Act: -0.0001

Data Index: 1127629 (C4 (Web Text))

Max Activating Token Index: 567

Text #13

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.7999. Min Act: -0.0001

Data Index: 1455933 (Python Code)

Max Activating Token Index: 301

Text #14

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.8140. Min Act: -0.0001

Data Index: 1412801 (Python Code)

Max Activating Token Index: 273

Text #15

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.7864. Min Act: -0.0001

Data Index: 1165208 (C4 (Web Text))

Max Activating Token Index: 364

Text #16

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.7882. Min Act: -0.0001

Data Index: 1447446 (Python Code)

Max Activating Token Index: 278

Text #17

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.7868. Min Act: -0.0001

Data Index: 350559 (C4 (Web Text))

Max Activating Token Index: 61

Text #18

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.7786. Min Act: -0.0001

Data Index: 888654 (C4 (Web Text))

Max Activating Token Index: 224

Text #19

Max Range: 1.0295. Min Range: -1.0295

Max Act: 0.6804. Min Act: -0.0001

Data Index: 56170 (C4 (Web Text))

Max Activating Token Index: 577